Innovation & product
Design Thinking
A human-centred approach to innovation that works from empathy with users through problem definition, ideation, prototyping and testing. Popularised by IDEO and Tim Brown, and taught worldwide through the Stanford d.school's five-stage model.
Also known as Human-centred design, The d.school five-stage model. First set out by IDEO and the Stanford d.school, popularised by Tim Brown; intellectual roots in Herbert A. Simon in 2008; the primary source is cited in full below.
Where this is contested
The five-stage model is the Stanford d.school's teaching scaffold, layered on a longer lineage from Simon's The Sciences of the Artificial (1969) through David Kelley's IDEO practice to Brown's 2008 article; many competing stage models exist and no single canonical process or originator can be credited.
- Format
- Process / loop
- Level
- Product · Team
- Best for
- Understand customers · Evaluate options
- Decision stage
- Diagnose · Explore options
- Difficulty
- Intermediate
- Time to apply
- One to six weeks for a full cycle; a first rough loop can run in days.
Plate · The model
The components
Empathise
Fieldwork to understand users and their context: observation, interviews and immersion. The foundation stage; everything downstream inherits its quality.
Signals of strength
Team members have watched real users in context, not just read reports · Notes record behaviour and emotion, not only stated preferences · Surprising or contradictory observations are captured rather than smoothed over
Define
Synthesis of research into an actionable point of view: a specific user, an unmet need and the insight behind it. The stage where the problem is chosen, which makes it the highest-leverage stage in the process.
Signals of strength
The problem statement names a user and a need, not a product feature · The definition reframes or sharpens the original brief · The team can say what evidence would prove the definition wrong
Ideate
Structured generation of many candidate solutions to the defined problem, deferring judgement before converging on a shortlist worth prototyping.
Signals of strength
Idea volume is high and includes genuinely unreasonable options · Selection criteria are agreed before ideas are culled · Ideas trace back to the point of view rather than to pet projects
Prototype
Cheap, fast, disposable artefacts that make ideas testable: sketches, mock-ups, role-plays, staged services. Built to learn, not to impress.
Signals of strength
Prototypes are built in hours or days, not weeks · Each prototype embodies a specific question · The team is visibly willing to throw prototypes away
Test
Putting prototypes in front of real users to observe behaviour, gather reactions and decide whether to refine, pivot or return to an earlier stage.
Signals of strength
Users interact with the prototype rather than being pitched it · Negative results change the plan instead of being explained away · Findings loop back into Define or Empathise when they contradict the framing
When it earns its keep
- The problem is poorly defined and you suspect the organisation is solving the wrong one. The Empathise and Define stages exist to reframe before anyone commits to a solution.
- You are designing a product, service or experience where user behaviour, not technical feasibility, is the main uncertainty.
- The team is anchored on one obvious answer and needs a disciplined way to generate and test genuine alternatives cheaply.
- Cross-functional groups need a shared process for working on a customer problem without defaulting to the loudest voice in the room.
And when it doesn't
- The problem is well defined and the constraint is execution. If everyone agrees what to build, prototyping the question again is theatre; move to delivery.
- The decisive uncertainty is technical, regulatory or economic rather than behavioural. Empathy interviews will not tell you whether the unit economics work; pair with break-even or unit-economics analysis.
- There is no budget or appetite to act on what users reveal. Running workshops with no route to implementation produces sticky notes and cynicism.
- You need a portfolio or corporate-strategy view. Design thinking works at the level of a specific user and problem, not the level of where to compete.
How to run it
Before starting, gather the inputs the analysis depends on:
- Access to real users in their own context, for observation and interviews, not just survey data about them.
- A cross-functional team with the authority to reframe the brief if the evidence demands it.
- Materials and time for rough prototyping, plus permission to show unfinished work to real users.
- A sponsor who accepts that the first defined problem may not be the one the project ends up solving.
- 1
Empathise with users in context
Observe and interview the people you are designing for where the problem actually occurs. The aim is to surface latent needs and workarounds that users cannot articulate in a survey. Notes should capture what people do and feel, not just what they say.
- 2
Define a point of view
Synthesise the research into a sharp, human-centred problem statement: a specific user, a real need, and the insight that explains it. A good definition usually reframes the original brief. If your problem statement could have been written before the fieldwork, the fieldwork has not been used.
- 3
Ideate widely before choosing
Generate a broad set of possible responses to the defined problem, deferring judgement, then converge on a few worth building. Quantity first, selection second. The stage fails when the team ideates towards a solution it had already chosen.
- 4
Prototype cheaply
Build the fastest, roughest artefact that lets a user experience the idea: a paper mock-up, a role-play, a staffed fake service. The prototype is a question made physical. Spend hours, not weeks; polish at this stage only slows learning and raises attachment.
- 5
Test with users and loop back
Put prototypes in front of real users, watch what they do, and treat surprises as data. Results feed back into any earlier stage: refine the prototype, redefine the problem, or return to the field. The model is drawn as a sequence but practised as a loop.
Reading the result
A validated problem definition grounded in user research, a tested prototype of the strongest response to it, and an evidence trail of what users actually did that either justifies investment or kills the idea cheaply.
- The most valuable output is often the redefined problem, not the prototype. If the point of view at the end matches the brief at the start, be suspicious that the process was run as ritual.
- Read test results behaviourally: what users did with the prototype outweighs what they said about it.
- A killed idea with clear evidence is a success. The process exists to make failure cheap and early rather than expensive and late.
A worked example
An airport lounge operator redesigns the experience for time-pressured travellers
A UK operator of pay-on-entry lounges at regional airports is losing repeat custom. Satisfaction scores are falling despite a recent catering upgrade, and the board is debating a further refurbishment. Before committing capital, the product team runs a design thinking cycle on the business-traveller segment.
- Empathise
- The team spends four days across two lounges observing guests and running short interviews at departure. The striking pattern is not about food or seating: guests repeatedly check departure boards, sit facing the exit, and leave far earlier than needed. The dominant emotion in the lounge is low-level boarding anxiety, not relaxation.
- Define
- Point of view: a business traveller paying for the lounge needs certainty about time-to-gate, because until they trust they will not miss boarding they cannot use anything else the lounge sells. This reframes the brief from 'improve the lounge offer' to 'remove boarding anxiety'.
- Ideate
- A cross-functional session generates around forty ideas, from live gate-walk timers and personal boarding alerts to a staffed flight desk, an escorted fast-track promise and a 'relax until we call you' guarantee. The shortlist is chosen for testability, not glamour: personal SMS gate alerts and a visible staffed flight-watch board.
- Prototype
- Both ideas are faked cheaply in one lounge within a week. A host with a tablet sends manual text alerts to opted-in guests, and a whiteboard by the buffet shows live walk-to-gate times maintained by staff. Total spend is under five hundred pounds.
- Test
- Over two weeks, opted-in guests stay in the lounge an average of eighteen minutes longer, secondary spend per head rises, and exit interviews show they trust a personal alert more than any screen. The flight-watch board is glanced at but does not change behaviour. One finding loops back: anxious guests want the alert to include a walking-time buffer they choose.
The read. The cycle kills the refurbishment as the answer to falling scores and redirects investment towards a boarding-assurance service: opt-in personal alerts with a guest-set buffer, staffed at peak. The catering upgrade was solving a problem guests did not have. The prototype evidence, not opinion, is what changed the capital plan.
Pitfalls
- Running the stages as a linear ritual. The model is a loop with returns; teams that refuse to revisit Define when tests contradict it are performing the process, not using it.
- Empathy theatre: a handful of friendly interviews used to decorate a solution that was already chosen.
- Prototyping at too high a fidelity, which raises cost, slows iteration and makes the team defend the artefact instead of learning from it.
- Workshopping without a route to delivery. Design thinking produces direction and evidence; without engineering, operations and funding downstream it produces wall art.
- Treating the five stages as the method itself. They are a teaching scaffold; the underlying discipline is abductive reasoning about user problems, which the scaffold can hide.
What the critics say
Iskander argues design thinking is fundamentally conservative: it privileges the designer over the people served, taming problem-solving into a tidy process that preserves existing arrangements while claiming to challenge them, and limiting genuine participation in defining the problem.
Iskander, N. (2018) 'Design Thinking Is Fundamentally Conservative and Preserves the Status Quo', Harvard Business Review, 5 September 2018.
Vinsel's critique from the history of technology is that design thinking over-promises, giving students and organisations 'creative confidence' without real capability, and that as a movement it is more about commercialisation and branding than about design or meaningful innovation.
Vinsel, L. (2018) 'Design Thinking Is a Boondoggle', The Chronicle of Higher Education, 21 May 2018.
Scholars of design note that 'design thinking' bundles at least two distinct traditions, the academic study of designerly cognition and a management consultancy offer, and that the popular five-stage model has a thin evidence base connecting its use to innovation outcomes.
Johansson-Sköldberg, U., Woodilla, J. and Çetinkaya, M. (2013) 'Design Thinking: Past, Present and Possible Futures', Creativity and Innovation Management, 22(2), pp. 121-146.
Sources and further reading
- Brown, T. (2008) 'Design Thinking', Harvard Business Review, 86(6), June 2008. ↗
- Brown, T. (2009) Change by Design: How Design Thinking Transforms Organizations and Inspires Innovation. New York: HarperBusiness.
- Simon, H. A. (1969) The Sciences of the Artificial. Cambridge, MA: MIT Press.
- Hasso Plattner Institute of Design at Stanford, 'An Introduction to Design Thinking: Process Guide'. ↗